The 32 OD Pure Germanium Overdrive pedal - design and build in words and pictures

[laurie]

One last block to go - where we mix the overdrive signal with dry signal and ensure the output impedance of the pedal is low.

This is done simply with the other half of the JRC4558D opamp:
.

output2.jpg (74.21 KiB) Viewed 4717 times

[laurie]

Once again C54 is to minimize radio frequency interference. Plus in this instance it does roll off the very high audio frequencies in case some "frizz" gets through from the overdrive section.

For a standard inverting opamp circuit like this the gain is the feedback resistor divided by the input resistor.

So for the overdrive signal there is a gain of 100/33 = 3
For the clean signal there is a gain of 100/4.7 = 20

The clean gain is much higher so that if you want, you can dial back the overall level and wind up the clean to get a boost with a hint of overdrive.

The feedback resistor is a potentiometer so that the overall output level can be adjusted. When fully counterclockwise the resistance is zero, and there is no output signal... 0/33 = 0, and 0/4.7 = 0

The output impedance of the JRC4558D is very low (around 100 Ohms). It is therefore ideal for driving whatever comes after it, so a separate output driver stage isn't required. This half of the JRC4558D is also the output driver.

[laurie]

The 3PDT switch sits on a separate circuit board. The switch moves fractionally every time it is stomped on, so it is good practice to isolate it from the main circuit board.

There are lots of ways to do this. Over the years I've developed an approach that works. A small circuit board containing the switch, the power LED, and a resistor to "short" the pedal input to ground in bypass mode:
.

SW2.jpg (174.05 KiB) Viewed 4703 times

[laurie]

In the position shown, the pedal is bypassed.

The LED, controlled by SW1A is not powered (off).

SW1B connects the pedal electronics input to ground through R4. The reason this is done is so that the input capacitor C11 doesn't accumulate random electrical charge. Not having R4 can lead to audible pops when the pedal is switched from bypass to active as the charge accumulated on C11 is discharged.

SW1C connects the input jack to the output jack, mechanically bypassing the pedal.

[laurie]

Here is the schematic end-to-end.

REMEMBER:
- THIS IS COPYRIGHT
- THIS CAN NOT BE USED FOR COMMERCIAL GAIN
- NO WARRANTY OF ANY KIND IS PROVIDED
.

32 OD full schematic2.jpg (192.42 KiB) Viewed 4699 times

[laurie]

A few notes:
- R55 is there to limit the output current of the pedal. The JRC4558D is happy driving into a short circuit, but it is good practice to have this resistor.
- C53 is the output coupling capacitor. Ensures that the DC bias level of the 32 OD doesn't mess with whatever it is plugged into.
- R56 is there to stop random charge accumulating on the output coupling capacitor when the pedal is in bypass (similarly to R4).
- The two halves of the JRC4558D are "DC coupled". There is no AC coupling capacitor between them because they use the same 4.5V reference.
- The AC coupling capacitor for the clean signal is 0.1uF, not 1uF like the others. Because of the impedances on either side of it, 0.1uF is fine here.
- D2 is to protect the pedal from reverse voltage supply and over-voltage supply.
- Voltages of all capacitors are 25V or greater. Generally electrolytic capacitors (polarised) are 25V and others are 50V or 63V.
- Connecting pin 1 of the DC jack to the "ring" of the input jack means the pedal is only powered when the input jack has a mono plug inserted (just like nearly every other pedal you own).

And...

I decided to make R21 variable (now RV21). This will give more flexibility to use different transistors, plus allow fine-tuning of the bias point if anyone wants to do it.

[Pepe]

Your approach with the 3PDT and the additional 10k resistor is very interesting.

[laurie]

OK... so I have a confession to make.

A lot of the explanation in the last 5 pages is just skimming the surface. Putting things in high-level terms.

What is actually happening at an electrical engineering level is a lot more detailed and mathematical.

For example, a transistor isn't as simple as shown. There are many mathematical models, here are a few:
https://ecee.colorado.edu/~bart/book/bo ... /ch5_6.htm

This starts to explain why germanium sounds different to silicon. If you look at the SPICE model, there are parameters that can vary widely between transistor types.

• BF Forward active current gain

• BR Reverse active current gain

• IS Transport saturation current

• CJE Base-emitter zero-bias junction capacitance

• CJC Base-collector zero-bias Junction capacitance

• VJE Base-emitter built-in potential

• VJC Base-collector built-in potential

• VAF Forward mode Early voltage

• VAR Reverse mode Early voltage

Mostly, for pedals, the high-level approach works well enough. But there are places (like the overdrive section) where more detailed modelling helps.

The phase-versus-frequency response hasn't been discussed (and won't be here):
https://en.wikipedia.org/wiki/Bode_plot ... se%20shift.

The absolute frequency response hasn't been discussed. It is measured to be:

• Upper 3dB point is 24kHz

• Lower 3dB point for overdrive signal path is 20Hz

• Lower 3dB point for clean signal path is 200Hz

The lower 3dB point being higher for the clean path is intentional - when clean is added it will take a bit of the bottom end out of the mixed signal. Just sounds better to my ear (and possibly to sclitheroe's - re the comment about "an eye to being able to tighten up the bass just a bit when desired").

Additionally, in high gain, high impedance circuits a thing called "thermal noise" becomes important.
https://en.wikipedia.org/wiki/Johnson%E ... uist_noise

Except in rare cases, the higher the resistance or impedance, the more thermal noise the component contributes. That's part of the "hiss" you hear in high gain circuits. So there is always a balance between wanting things to be low-current (ie. higher resistance for longer battery life) and low resistance/impedance to keep the noise low. A lot of the resistor values in the 32 OD have been selected to reach this balance.

Another thing is the stability of bias points. Take the R1-R2 voltage divider. It would be nice if we could make that out of high value resistors (say 1,000,000 Ohms) to reduce battery current through R1 and R2. But because the 4.5V reference will be sinking or sourcing current as the signal is processed the bias point will move around slightly with resistors that high. And that may not sound good. R1 and R2 need to allow current to flow through them that is (as a rule-of-thumb) is at least 10 times greater than any current that will be sourced or sunk on the 4.5V reference. About 4700 Ohms is about right.

4700 Ohms gives a stable bias point, and we have stayed away from the increased thermal noise that 1,000,000 Ohm resistors produce. But we have done it at the expense of 1mA of current draw from the battery.

And then there are decisions about "do you want any DC to flow through potentiometers"? If there is DC flowing in a potentiometer it can make a "dragging" sound when you move the knob. Trade that off against adding a capacitor into the signal path to block that DC (more "stuff" in the way of your tone).

Balance... designs are always always about balance.

What I'm hoping to do is to give an appreciation that there is more going on than meets-the-eye in our bus-ride past the design in the previous pages.

[laurie]

Laying out the circuit board comes next...

[sclitheroe]

So how thermally stable would a design like this be compared to say a germanium fuzz face? What would contribute to similar or differing characteristics in that regard?